A direct inversion scheme for deep resistivity sounding data using artificial neural networks

Abstract

Initialization of model parameters is crucial in the conventional 1D inversion of DC electrical data, since a poor guess may result in undesired parameter estimations. In the present work, we investigate the performance of neural networks in the direct inversion of DC sounding data, without the need of a priori information. We introduce a two-step network approach where the first network identifies the curve type, followed by the model parameter estimation using the second network. This approach provides the flexibility to accommodate all the characteristic sounding curve types with a wide range of resistivity and thickness. Here we realize a three layer feed-forward neural network with fast back propagation learning algorithms performing well. The basic data sets for training and testing were simulated on the basis of available deep resistivity sounding (DRS) data from the crystalline terrains of south India. The optimum network parameters and performance were decided as a function of the testing error convergence with respect to the network training error. On adequate training, the final weights simulate faithfully to recover resistivity and thickness on new data. The small discrepancies noticed, however, are well within the resolvability of resistivity sounding curve interpretations.

A robust behavior of Feed Forward Back propagation algorithm of Artificial Neural Networks in the application of vertical electrical sounding data inversion

“…The ANN inversion has been used for many years to interpret electrical resistivity data (e.g. Dey and Morrison, 1979;Smith and Vozoff, 1984;Tripp et al, 1984;Constable et al, 1987;Uchida and Murakami, 1990;Poulton et al, 1992;Griffith and Barker, 1993;Lake and Barker, 1996;Calderon-Macias et al, 2000;Vander Baan and Jutten, 2000;El-Qady and Ushijima, 2001;Stephen et al, 2004;Neyamadpour et al, 2010). ANNs are also being increasingly applied in the field of engineering geophysics (Brown and Poulton, 1996).…”

A novel and generalized approach in the inversion of geoelectrical resistivity data using Artificial Neural Networks (ANN)

“…These networks have self-learning capability and are fault-tolerant as well as noise-immune and have applications in various fields (Sreekanth et al 2009;Singh et al 2010;Maiti et al 2011). Some of the researchers experimented with artificial neural network tool in geophysical problems very recently (El Qady et al 2001;Jimmy Stephen et al 2004;Singh et al 2005;Tiwari 2008, 2009;Srinivas et al 2010Srinivas et al , 2012. However the ID resistivity inversion problem is successful only when the problem is subjected to a particular case study.…”

Improved Monte Carlo 1D inversion of vertical electrical sounding and time‐domain electromagnetic data

“…Furthermore, MCMC methods typically use only a priori fixed number of layers; also algorithms with varying number of layers have been developed with success (e.g., Malinverno ). Recently, also artificial neural networks (NN) application to resistivity data inversion has been developed (El‐Qady and Ushijima ; Stephen et al ). If adequately trained, NN produce reasonably accurate earth models (Calderón‐Macías et al ).…”